The hexokinase 2 protein participates in regulatory DNA-proteincomplexes necessary for glucose repression of the SUC2 gene

in Saccharomyces cerevisiae

Pilar Herrero, Carlos Mart|ènez-Campa, Fernando Moreno*Departamento de Bioqu|èmica y Biolog|èa Molecular, Instituto Universitario de Biotecnolog|èa de Asturias, Universidad de Oviedo, 33006 Oviedo, Spain

Received 15 June 1998

Abstract The HXK2 gene plays an important role in glucoserepression in the yeast Saccharomyces cerevisiae. Recently wehave described that the HXK2 gene product, isoenzyme 2 ofhexokinase, is located both in the nucleus and in the cytoplasm ofS. cerevisiae cells. In this work we used deletion analysis toidentify the essential part of the protein-mediating nuclearlocalisation. Determinations of fructose-kinase activity andimmunoblot analysis using anti-Hxk2 antibodies in isolatednuclei, together with observations of the fluorescence distributionof Hxk2-GFP fusion protein in cells transformed with anHXK2: :gfp mutant gene, indicated that the decapeptideKKPQARKGSM, located between amino acid residues 7 and16 of hexokinase 2, is important for nuclear localisation of theprotein. Further experimental evidence, measuring invertaseactivity in wild-type and mutant cells expressing a truncatedversion of the Hxk2 protein unable to enter the nucleus, showsthat a nuclear localisation of Hxk2 is necessary for glucoserepression signalling of the SUC2 gene. Furthermore, wedemonstrate using gel mobility shift analysis that Hxk2participates in DNA-protein complexes with cis-acting regula-tory elements of the SUC2 gene promoter.z 1998 Federation of European Biochemical Societies.

Key words: HXK2; SUC2; Glucose repression;Saccharomyces cerevisiae

1. Introduction

Genetic analysis of Saccharomyces cerevisiae has led to theidenti¢cation of several genes necessary for glucose repressionand for derepression of enzyme synthesis after depletion ofglucose. Several gene products (Hxk2, Grr1, Reg1, Glc7,Mig1, Ssn6, Tup1) act negatively in glucose repression of di-saccharide utilising enzymes (for reviews see [1^3]). The SUC2gene of S. cerevisiae encodes the secreted enzyme invertasewhich hydrolyses sucrose and ra¤nose. Like other genes re-quired for utilisation of alternative carbon sources, expressionof the SUC2 gene is repressed by glucose. A central compo-nent and also one of the ¢rst gene products acting in theSUC2 glucose repression cascade seems to be HXK2 [4], butit is not known how glucose modulates Hxk2 functionthrough these proteins. It has been proposed that the sugar-kinase activity of hexokinase 2 is correlated with glucose re-pression [5,6]. However, if the glucokinase gene (GLK1) isoverexpressed in a hexokinase 1/hexokinase 2 double-null mu-tant genetic background no e¡ect on glucose repression isobserved, even in strains with a threefold increase of sugar-

phosphorylating activity [6]. This indicates that glucose re-pression is not only associated with the sugar-kinase activityof hexokinase 2 but that the presence of the Hxk2 protein isalso necessary to give the signal for glucose repression. Re-cently, it has been described that an HXK2 mutation encodinga serine to alanine change at position 15 produces a Hxk2protein unable to undergo phosphorylation [7] and that thetransformed cells with the HXK2(S15A) mutant gene couldnot provide glucose repression of invertase, suggesting thatthe phosphorylation of Hxk2 is essential in vivo for glucosesignal transduction [8].

On the other hand, it has also been described that the Hxk2is located in both the nucleus and the cytoplasm of S. cerevi-siae cells [9] and this observation opens new possibilities to-ward explaining the role of Hxk2 in glucose repression signal-ling.

Here we report that the nuclear localisation of Hxk2 pro-tein is mediated by an internal decapeptide sequence identi¢edas a nuclear localisation sequence (NLS) that is necessary todirect the protein to the nucleus. We also show that a nuclearlocalisation of the Hxk2 protein is necessary for glucose re-pression signalling of the SUC2 gene. Furthermore, we showthat the nuclear Hxk2 is involved in the formation of speci¢cDNA-protein complexes during glucose-induced repression ofthe SUC2 gene. These results suggest that the Hxk2 proteincould participate in the transduction of the glucose repressionsignal by interacting with transcriptional factors related to thisregulatory mechanism.

2. Materials and methods

2.1. Strains and plasmidsS. cerevisiae strains DBY1315 (MATK ura3-52 leu2-3,2-112 lys2-801

gal2) and DBY2052 (MATK hxk1: :LEU2 hxk2-202 ura3-52 leu2-3,2-112 lys2-801 gal2) were donated by D. Botstein, and were used asrecipients in transformation experiments. Bacterial transformationand large-scale preparation of plasmid DNA were performed in Es-cherichia coli MC1061.

Plasmids YEp356 [10] and pRS306 [11] are yeast-E. coli shuttlevectors.

2.2. Media, growth conditions and enzymatic analysisYeasts were grown on 1% yeast extract and 2% peptone supple-

mented with 4% glucose (YEPD). The cells were grown in this me-dium until the optical density at 600 nm reached 1.0. To select fortransformants, synthetic medium with yeast nitrogen base, 2% glucoseand the adequate supplements was used.

Invertase and hexokinase were assayed as described by Moreno etal. [12]. The protein concentrations were determined according to [13],using bovine serum albumin as the standard. Speci¢c activities areexpressed as nmol substrate consumed/min/mg protein.

2.3. Preparation of crude protein extractsYeast protein extracts were prepared as follows: yeast was grown

FEBS 20634 27-8-98 Cyaan Magenta Geel Zwart

0014-5793/98/$19.00 ß 1998 Federation of European Biochemical Societies. All rights reserved.PII: S 0 0 1 4 - 5 7 9 3 ( 9 8 ) 0 0 8 7 2 - 2

*Corresponding author. Fax: (34) (98) 510 3157.E-mail: fms@sauron.quimica.uniovi.es

FEBS 20634 FEBS Letters 434 (1998) 71^76

on 10^20 ml of rich medium (YEPD) at 28³C until an optical densityat 600 nm of 1.0. Cells were collected, washed twice with 1 ml of 1 Msorbitol and suspended in 100 Wl 50 mM Tris-HCl (pH 7.5) bu¡ercontaining 0.2 mM EDTA, 0.5 mM dithiothreitol (DTT), 0.5 mMphenylmethylsulphonyl £uoride (PMSF), 0.42 M NaCl and 1.5 mMMgCl2. The cells were broken by vortexing (6U20 s) in the presenceof glass beads (0.5 g), and 400 Wl of the same bu¡er was added to thesuspension. After centrifugation at 19 000Ug (14 000 rpm) for 15 minat 4³C, the supernatant was used as crude protein extract.

2.4. General DNA techniquesRestriction enzymes and T4 DNA ligase were from Boehringer,

Sequenase V2.0 from USB. Radioactively labelled isotopes werefrom Amersham International. The dideoxyribonucleotide chain ter-mination procedure was used for DNA sequencing analysis [14]. Allother DNA manipulations were as previously described [15].

2.5. Construction of yeastY strains with HXK2, HXK2(S15A),HXK2vK7 M16 , HXK2: :gfp or HXK2vK7 M16 : :gfp genes

A DNA fragment containing the complete HXK2 promoter wasisolated from the vector pRS-HXK2 [16] as a 0.88 kb SphI-NcoIfragment and subcloned into a SphI-NcoI previously cleaved vectorpSP73-HG (this plasmid contains in a 2.75 kb fragment the completecoding region of HXK2 gene and 254 bp of the 5P non-coding region).The resulting plasmid pSP73-HXK2 contains in a 3.35 kb SphI-EcoRIfragment the complete HXK2 gene under the control of its own pro-moter. This fragment was cloned into YEp356. The resulting plasmidwas called YEp356-HXK2.

In vitro mutagenesis of HXK2 was done using the Sculptor muta-genesis system (Amersham). The template used for mutagenesis wasbacteriophage M13mp18, with an inserted 750 bp Asp718 fragment ofthe HXK2 gene and promoter obtained from pSP73-HXK2. The nu-cleotide 5P-AAAGGGTgccATGGCCG-3P was used in the mutagene-sis, the changed codon is shown in lower case. The 750 bp Asp718mutated fragment from a positive phage was subcloned in YEp356-HXK2 cleaved with the same enzymes. The resulting plasmidYEp356-HXK2(S15A) carries a HXK2 allele codifying a hexokinase2 with Ser15 changed to Ala.

Oligonucleotides 5P-ATGAACCATTTTATT-3P (primer 1) and 5P-GGTTCCATGGCCGATGTG-3P (primer 2) were used to generateconstruction pSP73-HXK2vK7M16 by PCR. The construct was ob-tained by using 0.1 Wg of the primer pair 1+2, 1 Wg of pSP73-HXK2as template, 2.5 U of Taq Polymerase (Promega), 0.2 mM dNTPs(Pharmacia) in a total reaction volume of 25 Wl in reaction bu¡erprovided by the manufacturer for 30 cycles at 94³C for 30 s, 55³Cfor 90 s and 72³C for 60 s. The PCR product was isolated from 0.8%agarose gels and ligated. The resulting plasmid (pSP73-HXK2vK7M16) was cleaved by SphI and EcoRI to obtain an approx-imately 3.35 kb fragment which was subcloned into a SphI-EcoRIpreviously cleaved vector YEp356. The plasmid obtained, YEp356-HXK2vK7M16, has a 30 nucleotides deletion between nucleotides+19 and +48 of the HXK2 gene. Expression of this mutant genegenerates a truncated Hxk2 protein with similar speci¢c activity asthe wild-type hexokinase 2 but lacking amino acids from K7 to M16.

A 969 bp PstI-BglII fragment containing the gfp gene was sub-cloned into the pSP73-HXK2 and the pSP73-HXK2vK7M16 vectors,¢rst cleaved with PstI-BglII. The resulting plasmids pSP73-HXK2: :gfp and pSP73-HXK2vK7M16 : :gfp were used to obtainXhoI-BglII fragments containing respectively the HXK2 gene andthe HXK2vK7 M16 mutant gene fused in frame with the gfp gene.These fragments were subcloned in a pRS306 plasmid ¢rst cleavedwith XhoI-BamHI revealing plasmids pRS306-HXK2: :gfp andpRS306-HXK2vK7M16 : :gfp.

All the clones used were veri¢ed by sequencing analysis of fusionpoints.

Plasmids YEp356-HXK2, YEp356-HXK2(S15A) and YEp356-HXK2vK7M16 were used to transform the yeast strain DBY2052and plasmids pRS306-HXK2: :gfp and pRS306-HXK2vK7M16 : :gfpwere used to transform yeast strain DBY1315.

2.6. Preparation of yeast nuclei and nuclear extractsNuclei were prepared from strains DBY1315; DBY2052; DBY1315

transformed with plasmids pRS306-HXK2: :gfp or pRS306-HXK2vK7M16 : :gfp; and DBY2052 transformed with plasmidsYEp356-HXK2, YEp356-HXK2(S15A) or YEp356-HXK2vK7M16,

by the method described previously [9]. Nuclear extracts were pre-pared as follows: the band, consisting of clean nuclei [9], collectedfrom near the top of the Percoll gradient was diluted three times withlysis bu¡er (50 mM Tris-HCl, pH 7.5; 10 mM Mg2SO4 ; 1 mMEDTA; 10 mM potassium acetate; 1 mM DTT and 1 mM PMSF),and centrifuged at 19 000Ug for 10 min at 4³C. The supernatant wascarefully removed by aspiration and the nuclei pellet was resuspendedin 100 Wl of lysis bu¡er. The nuclei were broken by vortexing (6U20 s)in the presence of glass beads (0.5 g), and 200 Wl of the same bu¡erwas added to the suspension. After centrifugation at 19 000Ug for15 min at 4³C, the supernatant was used as nuclear extract.

2.7. DNA probesOligonucleotides, corresponding to both strands of the UAS ele-

ments (underlined) of the SUC2 gene [17,18], were synthesised withan added TCGA nucleotide overhang at the 5P-terminal end.OL34SUC2 sense: 5P-tcgaGTTTAGGAAATTATCCGGGGGCGAA-GAAATACGC-3P ; OL34SUC2 antisense: 5P-tcgaGCGTATTTCTT-CGCCCCCGGATAATTTCCTAAAC-3P.

The complementary strands were annealed and either end-labelledwith [K-32P]dCTP using the Klenow fragment of DNA polymerase Ior used as unlabelled competitors in protein-binding experiments.

2.8. Gel retardation assaysBinding reactions mixtures contained 10 mM HEPES (pH 7.5),

1 mM DTT, 1^5 Wg of poly(dI-dC) and 0.5 ng of end-labelledDNA in a volume of 25 Wl. When unlabelled competitor DNA wasadded, its amount is indicated in the ¢gure legends. The bindingreaction mixtures included 12 Wg (6 Wl) of protein from a nuclearextract and after 30 min of incubation at room temperature theywere loaded onto a 4% non-denaturing polyacrylamide gel. Electro-phoresis was carried out at 10 V/cm of gel for 45 min to 1 h in0.5UTBE bu¡er (45 mM Tris-borate, 1 mM EDTA). Gels were driedand autoradiographed at 370³C with an intensifying screen.

2.9. Electrophoretic analysis, immunoblotting and antibodiesElectrophoresis of proteins (SDS-PAGE) was performed on 10%

FEBS 20634 27-8-98 Cyaan Magenta Geel Zwart

Table 1Sequence present within the amino-terminal 16 residues of the Hxk2protein and in the other known or presumed nuclear proteins

MatK1 and MatK2 [32], H2B.1 [33], Gal4 [34], H4 [35], H2A.1 [36],Spt2 [37], Hex2 [38], Ski3 [39], Rad52 [40] and Swi5 [41]. Numbersbelow an amino acid indicate the position of the amino acid withinthe respective protein. Numbers in parentheses indicate the totalnumber of amino acids in the designated proteins.*Consensus sequence also present at positions 595 and 875.

P. Herrero et al./FEBS Letters 434 (1998) 71^7672

polyacrylamide gels using the bu¡er system described in [19]. Westerntransfer of proteins to a nitrocellulose membrane was carried out asdescribed in [20]. Hxk2 protein was detected by sequential incubationwith crude polyclonal antibody (1:1500 dilution) and protein A-per-oxidase (1:4000 dilution). Speci¢c anti-Hxk2 serum was raised inrabbits by sequential immunisation with a puri¢ed fraction of hexo-kinase 2 [21].

3. Results

3.1. Identi¢cation of a NLS in the Hxk2 proteinPrevious results obtained by measuring hexokinase activity

and by detecting Hxk2 protein with speci¢c antibodies in iso-lated nuclei indicate a double cytosolic-nuclear localisation ofthe Hxk2 protein in yeast growing in glucose-containing me-dia. The localisation of a fraction of the Hxk2 protein in thenucleus was further con¢rmed by expressing a Hxk2-GFPfusion protein in yeast ruling out a possible cross-contamina-tion during subcellular fractionation [9]. Therefore we furtherinvestigated the presence in Hxk2 of signal sequences de-scribed previously as nuclear-targeting signals.

A large number of NLSs have been characterised to dateand that allows analysis of their general features. All containa number of basic residues, but they do not conform to theconsensus bipartite sequence proposed previously [22]. How-ever, comparison of amino acids sequences of other known orpresumed yeast nuclear proteins with the amino-terminal 16residues of Hxk2 reveals a sequence that might be importantfor nuclear targeting, Lys8-Pro-Gln-Ala-Arg12 (Table 1). Asimilar sequence of two positively charged amino acids £ank-ing three residues, one of which is proline, is present in severalother yeast nuclear proteins but this sequence is not present inany yeast cytoplasmic proteins currently known [22,23].

The nuclear targeting ability of this putative NLS identi¢edin Hxk2 was determined by deletion analysis. A HXK2 mu-tant gene (HXK2vK7 M16 ) was constructed by PCR, and thisgene has a 30 bp deletion between nucleotides +19 and +48.The expression of this mutant gene generates a truncatedHxk2 protein with similar speci¢c activity as the wild-typehexokinase 2 but without the amino acids from Lys7 to Met16.

As is shown in Table 2, the wild-type phenotype was par-tially restored after transformation of a hxk1/hxk2 doublemutant strain with a HXK2vK7 M16 -containing plasmid. Themutant strain produced about 97% of the enzyme activitycompared to that of wild-type cells, but no hexokinase activitywas detected in isolated nuclei from this strain. However, thewild-type phenotype was completely restored after transfor-mation of hxk1 hxk2 double mutant strain with theYEP356-HXK2 or the YEP356-HXK2(S15A) plasmids. More-

over, laser scanning confocal microscopy (Fig. 1) demon-strated that Hxk2: :GFP and Hxk2vK7M16 : :GFP fusion pro-teins were distributed in the cell in a manner consistent with asingle cytosolic localisation, such that each cell had an uni-form distribution of green £uorescence except the vacuolewhich was virtually free (Fig. 1a,d). Furthermore, nuclei pu-ri¢ed by Percoll gradients from wild-type cells transformedwith a plasmid containing the HXK2: :gfp gene showed aclear associated £uorescence (Fig. 1c). However, nuclei puri-¢ed by Percoll gradients from wild-type cells transformed witha plasmid containing the HXK2vK7 M16 : :gfp gene do notshow any associated green £uorescence (Fig. 1f).

These results were con¢rmed by immunoblot analysis usingpolyclonal antibody against Hxk2 protein. As we can see inFig. 2, Hxk2 protein was detected in the crude extracts ob-

FEBS 20634 27-8-98 Cyaan Magenta Geel Zwart

Fig. 1. Detection of Hxk2-GFP fusion protein in living yeast cells.Yeast strain DBY1315 was transformed with the integrative expres-sion plasmids pRS306-HXK2: :gfp or pRS306-HXK2vK7M16 : :gfp.The resulting single copy transformed strains were grown exponen-tially in YEPD liquid media and cell-associated £uorescence was an-alysed in whole cells transformed with the pRS306-HXK2: :gfp plas-mid (a) or the pRS306-HXK2vK7M16 : :gfp plasmid (d) by confocalmicroscopy. Panels b and e are corresponding nuclei photographedby phase contrast. The £uorescence associated with the nuclei iso-lated from cells transformed with the pRS306-HXK2: :gfp plasmid(c) or the pRS306-HXK2vK7M16 : :gfp plasmid (f) was also analysedby confocal microscopy. Confocal images of Hxk2-GFP fusion pro-tein expression were obtained on a Bio-Ras MRC 600 inverted laserconfocal microscope using a standard £uorescein isothiocyanate ¢l-ter providing excitation at 490 nm and emission at 527 nm. The im-age ¢les were processed using a computer-based graphic system(COMOS) where they were arranged and annotated. (a, d: U150;b, e: U1000; c, f: U2000).

Table 2Speci¢c hexokinase activity in crude extracts and nuclear extracts from di¡erent yeast strains

Strain Plasmid Hexokinase (mU/mg protein)

Crude extract Nuclear extract

DBY1315 ^ 1150 161DBY2052 ^ n.d. n.d.DBY2052 YEp356 n.d. n.d.DBY2052 YEp356-HXK2 1108 153DBY2052 YEp356-HXK2vK7M16 1120 n.d.DBY2052 YEp356-HXK2(S15A) 1109 166

Plasmids YEp356-HXK2, YEp356-HXK2(S15A) and YEp356-HXK2vK7M16 are described in Section 2. Transformed cells with plasmid YEp356were used as reference. Cells were grown in YEPD medium, harvested at the mid-log phase (A600, 1.0) and fractionated as described in Section 2.The hexose-phosphorylating activity was measured using fructose as substrate. n.d., not detectable.

P. Herrero et al./FEBS Letters 434 (1998) 71^76 73

tained from all the yeast strains analysed, except from thehxk1 hxk2 double mutant strain as expected. However,Hxk2 was found in the nuclear extract obtained from thewild-type strain and the double mutant strain transformedwith YEP356-HXK2(S15A) plasmid, but not in the nuclearextract obtained from the double mutant strain or the doublemutant strain transformed with a HXK2vK7 M16 -containingplasmid.

Therefore, we can conclude that the accumulation of hexo-kinase 2 in the nuclear fraction strongly depends on the pres-ence of a NLS, which was localised between amino acids Lys7

and Met16 of the Hxk2 protein.

3.2. The nuclear localisation of Hxk2 is required for glucoserepression of SUC2 gene

A number of authors have proposed signalling pathwaymodels in order to integrate positively and negatively actingfactors into one comprehensive scheme [1^3,24,25]. Thesemodels account for the following details : the presence of glu-cose must be sensed in the cells; this information is trans-duced by intracellular messengers and speci¢c target genesare ultimately turned on or o¡. Assuming that Hxk2 is anegative regulatory element and that a fraction of this proteinis localised inside the nucleus during the repression cycle, itshould be possible that the Hxk2 protein may interact withpositively or negatively acting factors involved in the glucoserepression cascade. Thus, our hypothesis assumes that: ¢rst,the expression of the HXK2 gene acts as sensor of the glucoseconcentration in the culture medium [26], second, a nuclear

localisation of the Hxk2 protein is required for glucose repres-sion signalling.

Data presented in Table 3 clearly show that the wild-typephenotype (low invertase activity in cells growing in repressingmedium) was restored after transformation of the hxk1 hxk2double mutant with the HXK2 or the HXK2(S15A) genes.When the hxk1 hxk2 double mutant was transformed withthe NLS mutant gene (HXK2vK7 M16) the wild-type pheno-type was not restored. Moreover, a two-fold increase of exo-cellular invertase activity, compared to that of hxk1 hxk2double mutant cells, was observed, indicating an overexpres-sion of the SUC2 gene in this genetic background. Theseresults revealed that nuclear localisation of Hxk2 is essentialfor glucose repression signalling of the SUC2 gene.

As a ¢rst step to characterise the putative role as repressorof Hxk2, its capacity to participate in DNA-protein com-plexes necessary for glucose repression of SUC2 gene wastested. When a double stranded oligonucleotide (OL34SUC2),including the UAS sequences of the SUC2 gene [17,18], wassubjected to gel mobility shift analysis using nuclear extractsobtained from wild-type and mutant strains as protein sour-ces, we observed two di¡erent protein-DNA complexes (CIand CII) with nuclear extracts prepared from glucose-grownwild-type cells (Fig. 3, lane 2). Competition assays with a non-labelled oligonucleotide indicated that the binding was speci¢cin all cases (complex I was only partially displaced by the non-labelled oligonucleotide concentrations used). Regarding thee¡ects of hxk1 hxk2 double mutations on the formation of thecomplexes, it can be seen in Fig. 3 (lane 6) that complex CIwas not formed when nuclear extracts from this strain were

FEBS 20634 27-8-98 Cyaan Magenta Geel Zwart

Fig. 3. Gel retardation assays with oligonucleotides which containthe sequence from the UAS controlling regions of SUC2 gene. Gelmobility shift assays were performed with 32P-labelled OL34SUC2

and nuclear extracts from glucose growing yeast cells of the indi-cated strains, prepared as described in Section 2. Lane 1, no proteinadded; lanes 2^5, protein from nuclear extracts obtained from wild-type repressed cells (lane 5: with 2 Wl of anti-Hxk2 antibodies); lane6, protein from nuclear extracts obtained from hxk1 hxk2 doublemutant repressed cells; lane 7, protein from nuclear extracts ob-tained from hxk1 hxk2 double mutant repressed cells transformedwith plasmid YEp356-HXK2(S15A); lane 8, protein from nuclearextracts obtained from hxk1 hxk2 double mutant repressed cellstransformed with plasmid YEp356-HXK2vK7M16. Strains: WT.wild-type (DBY1315); DM, double mutant hxk1 hxk2 (DBY2052);vK7M16, DBY2052 transformed with plasmid YEp356-HXK2vK7M16 ; S15A, DBY2052 transformed with plasmid YEp356-HXK2(S15A).

Table 3Speci¢c activity of secreted invertase from di¡erent strains

Strain Plasmid Invertase(mU/mg protein)

DBY1315 ^ 11DBY2052 ^ 2800DBY2052 YEp356 2800DBY2052 YEp356-HXK2 12DBY2052 YEp356-HXK2vK7M16 5700DBY2052 YEp356-HXK2(S15A) 13

Plasmids YEp356-HXK2, YEp356-HXK2(S15A) and YEp356-HXK2vK7M16 are described in Section 2. Transformed cells withplasmid YEp356 were used as reference. The external invertase activ-ity was assayed in whole cells grown in YEPD medium and harvestedat the mid-log phase (A600, 1.0).

Fig. 2. Immunoblot analysis of the nuclear fraction from di¡erentstrains. Crude and nuclear extracts from glucose growing yeast cellsof the indicated strains, prepared as described in Section 2, werefractionated by SDS-PAGE, the proteins were transferred to nitro-cellulose membrane and Hxk2 protein was detected with polyclonalanti-Hxk2 antibodies. Strains: WT, wild-type (DBY1315); DM,double mutant hxk1 hxk2 (DBY2052); vK7M16, DBY2052 trans-formed with plasmid YEp356-HXK2vK7M16 ; S15A, DBY2052transformed with plasmid YEp356-HXK2(S15A). CE, crude extract;NE, nuclear extract.

P. Herrero et al./FEBS Letters 434 (1998) 71^7674

used. The wild-type phenotype was restored (Fig. 3, lane 7)when we use nuclear extracts from repressed cells of a hxk1hxk2 double mutant strain transformed with a plasmid con-taining the mutant gene HXK2(S15A). However, the wild-typephenotype was not restored when we used nuclear extractsfrom repressed cells of a hxk1 hxk2 double mutant straintransformed with a plasmid containing the NLS mutantgene HXK2vK7 M16 , as can be seen in Fig. 3 (lane 8).

To con¢rm that Hxk2 is present in the retarded complexesobtained with nuclear extracts from the wild-type strain, weuse a polyclonal anti-Hxk2 serum [21]. As can be observed inFig. 3 (lane 5) the anti-Hxk2 antiserum shifted the position ofthe CI protein-DNA complex towards the top of the gel, to amore slowly migrating complex. In the absence of proteinextract the anti-Hxk2 antiserum did not produce any complex(data not shown).

Together, these results allow us to conclude that the Hxk2is a component of at least one (CI) of the retarded complexesthat form the UAS elements of the SUC2 promoter withtrans-acting regulatory factors involved in the signalling ofthe glucose repression cascade.

4. Discussion

These studies demonstrate that one region of yeast hexoki-nase 2 is important in determining its nuclear localisation andits function as mediator of glucose repression. Deletion ofHxk2 from residues Lys7 to Met16 abolished both nuclearlocalisation and glucose repression, thus indicating the impor-tance in nuclear targeting of these amino acids of the NH2-terminal region of the protein.

In this paper, we provide evidence supporting a novel idea.When there is glucose in the medium, the HXK2 gene is ex-pressed and a fraction of the Hxk2 protein, about 14% oftotal hexokinase 2 activity, is found in the nuclear compart-ment [9]. A clear correlation between nuclear location of theprotein and glucose repression of the SUC2 gene has beendemonstrated. Deletion analysis identi¢ed the essential partof the protein-mediating nuclear localisation and glucose re-pression between amino acid residues 7 and 16 (Tables 2 and3; Fig. 2). To gain insight into possible functions of the Hxk2protein in the nucleus, we used gel mobility shift analysis. Thisanalysis revealed that Hxk2 participates directly as part of acomplex involved in glucose repression signalling of the SUC2gene (Fig. 3). At present, a DNA-binding function of theHxk2 protein does not seem to be probable. As a preliminaryhypothesis, we suppose nuclear Hxk2 to be a competitor fortranscriptional factors, thus preventing transcription of theSUC2 gene. This preliminary hypothesis raises the questionof how Hxk2 protein localisation is regulated in response toglucose. One possibility is that Hxk2 localisation may be regu-lated by phosphorylation.

Although the targeting role of nuclear localisation signalshas been known for some time, more recent results indicatethat NLS-dependent nuclear protein import is precisely regu-lated. Phosphorylation appears to be the main mechanismcontrolling the nuclear transport of a number of proteins,including transcription factors such as NF-UB, c-rel, dorsal,and SWI5 from yeast [27]. Even nuclear localisation of thearchetypal NLS-containing simian virus 40 large tumour anti-gen is regulated by phosphorylation of a `CcN' motif close tothe NLS [28]. The regulation of nuclear transport through

phosphorylation appears to be common in eukaryotic cellsfrom yeast and plants to higher mammals. Yeast hexokinase2 is known to be a phosphoprotein in vitro [29] and in vivo[30]. The principal and perhaps sole site of phosphorylationwas identi¢ed as Ser15 by mutation to alanine, which preventsphosphorylation in vitro [31] and in vivo [8]. Because Hxk2 isphosphorylated at a site close to or inside NLS, it should beinteresting to examine if Ser15 phosphorylation speci¢callyregulate NLS function. With this aim we constructed the mu-tant gene HXK2(S15A). Expression of this gene generates amutant Hxk2 protein with similar speci¢c activity as the wild-type hexokinase 2 but with a change at residue 15 from serineto alanine. Transformation of a hxk1 hxk2 double mutantstrain with a HXK2(S15A)-containing plasmid restores bothnuclear localisation of Hxk2 and glucose repression of SUC2gene (Tables 2 and 3; Fig. 2). These results suggest that phos-phorylation of Ser15 does not a¡ect the nuclear targeting abil-ity of the NLS. Thus, the Hxk2(S15A) protein can be trans-ported to the nucleus (Fig. 2, Table 2) and in our hands, thisnuclear localisation generates the glucose repression signallingof the SUC2 gene (Fig. 3).

We have shown that yeast hexokinase 2 combines featureswhich have been previously identi¢ed as important in nuclearprotein targeting: (i) a speci¢c localisation sequence, and (ii)speci¢c binding to DNA cis-acting regulatory complexes in-volved in control of gene expression. Further elucidation ofthe mechanism controlling nuclear transport will be necessaryto understand how these features function together.

Acknowledgements: We are very grateful to Dr. Jane Mellor for pro-viding the gfp gene and Dr. Colin R. Goding for critical reading ofthe manuscript. This work was supported by Grant PB94-0091-C02-02 from the DGICYT.

References

[1] Gancedo, J.M. (1992) Eur. J. Biochem. 206, 297^313.[2] Trumbly, J.R. (1992) Mol. Microbiol. 6, 15^21.[3] Ronne, H. (1995) Trends Genet. 11, 12^17.[4] Entian, K.D., Kopetzki, E., Froëhlich, K.U. and Mecke, D.

(1984) Mol. Gen. Genet. 198, 50^54.[5] Ma, H., Bloom, L.M., Walsh, C.T. and Botstein, D. (1989) Mol.

Cell. Biol. 9, 5643^5649.[6] Rose, M., Albig, W. and Entian, K.D. (1991) Eur. J. Biochem.

199, 511^518.[7] Kriegel, T.M., Rush, J., Vojtek, A.B., Clifton, D. and Fraenkel,

D.G. (1994) Biochemistry 33, 148^152.[8] Randez-Gil, F., Sanz, P., Entian, K.D. and Prieto, J.A. (1998)

Mol. Cell. Biol. 18, 2940^2948.[9] Randez-Gil, F., Herrero, P., Sanz, P., Prieto, J.A. and Moreno,

F. (1998) FEBS Lett. 425, 475^478.[10] Myers, A.M., Tzagolo¡, A., Kinney, D.M. and Lusty, C.J.

(1986) Gene 45, 299^310.[11] Sikorski, S. and Hieter, P. (1989) Genetics 122, 19^27.[12] Moreno, F., Fernaèndez, T., Fernaèndez, R. and Herrero, P. (1986)

Eur. J. Biochem. 161, 565^569.[13] Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J.

(1951) J. Biol. Chem. 193, 265^275.[14] Sanger, F., Nicklen, S. and Coulson, S.A. (1977) Proc. Natl.

Acad. Sci. USA 74, 5463^5467.[15] Herrero, P., Ram|èrez, M., Mart|ènez-Campa, C. and Moreno, F.

(1996) Nucleic Acids Res. 24, 1822^1828.[16] Mart|ènez-Campa, C., Herrero, P., Ram|èrez, M. and Moreno, F.

(1996) FEMS Lett. 137, 69^74.[17] Sarokin, L. and Carlson, M. (1986) Mol. Cell. Biol. 6, 2324^

2333.[18] Bu, Y. and Schmidt, M.C. (1998) Nucleic Acids Res. 26, 1002^

1009.

FEBS 20634 27-8-98 Cyaan Magenta Geel Zwart

P. Herrero et al./FEBS Letters 434 (1998) 71^76 75

[19] Laemmli, U.K. (1970) Nature 227, 680^685.[20] Towbin, H., Staehelin, T. and Gordon, J. (1979) Proc. Natl.

Acad. Sci. USA 76, 4350^4354.[21] Herrero, P., Fernaèndez, R. and Moreno, F. (1989) J. Gen. Mi-

crobiol. 135, 1209^1216.[22] Dingwall, C. and Laskey, R.A. (1991) Trends Biochem. Sci. 16,

478^481.[23] Osborne, M.A. and Silver, P.A. (1993) Annu. Rev. Biochem. 62,

219^254.[24] Neigeborn, L. and Carlson, M. (1987) Genetics 115, 247^253.[25] Erickson, J.R. and Johnston, M. (1994) Genetics 136, 1271^1278.[26] Herrero, P., Gal|èndez, J., Ruiz, N., Mart|ènez-Campa, C. and

Moreno, F. (1995) Yeast 11, 137^144.[27] Da, J. and Hubner, S. (1996) Physiol. Rev. 76, 651^685.[28] Xiao, C.Y., Hubner, S., Elliot, R.M., Caon, A. and Jans, D.A.

(1996) J. Biol. Chem. 271, 6451^6457.[29] Fernaèndez, R., Herrero, P., Fernaèndez, M.T. and Moreno, F.

(1986) J. Gen. Microbiol. 132, 3467^3472.[30] Vojtek, A.B. and Fraenkel, D.G. (1990) Eur. J. Biochem. 190,

371^375.[31] Kriegel, T.M., Rush, J., Vojtek, A.B., Clifton, D. and Fraenkel,

D.G. (1994) Biochemistry 33, 148^152.

[32] Astell, C.R., Ahlstrom-Jonasson, L., Smith, M., Tatchell, K.,Nasmyth, K.A. and Hall, B.D. (1981) Cell 27, 15^23.

[33] Wallis, J.W., Hereford, L. and Grunstein, M. (1980) Cell 22,799^805.

[34] Laughon, A. and Gesteland, R.F. (1984) Mol. Cell. Biol. 4, 260^267.

[35] Smith, M.M. and Andresson, O.S. (1983) J. Mol. Biol. 169, 663^690.

[36] Choe, J., Kolodrubetz, D. and grunstein, M. (1982) Proc. Natl.Acad. Sci. USA 79, 1484^1487.

[37] Roeder, G.S., Beard, C., Smith, M. and Keranen, S. (1985) Mol.Cell. Biol. 5, 1543^1553.

[38] Niederacher, D. and Entian, K.D. (1991) Eur. J. Biochem. 200,311^319.

[39] Sikorski, R.S., Boguski, M.S., Goebl, M. and Hieter, P. (1990)Cell 60, 307^317.

[40] Adzuma, K., Ogawa, T. and Ogawa, H. (1984) Mol. Cell. Biol. 4,2735^2744.

[41] Stillman, D.J., Bankier, A.T., Seddon, A., Groenhout, E.G. andNasmyth, K.A. (1988) EMBO J. 7, 485^494.

FEBS 20634 27-8-98 Cyaan Magenta Geel Zwart

P. Herrero et al./FEBS Letters 434 (1998) 71^7676